AIR SUSPENSIONS: DYNAMIC PERFORMANCE
John Lambert, Arnold McLean and BoHao Li
1
2
John Lambert and Associates Pty Ltd
Faculty of Engineering University of Wollongong.
ABSTRACT
In 1999 the author was requested to assist in evaluating a new air suspension developed by Bill Haire,
Director of Haire Truck and Bus Repairs Pty Ltd.
Initial theoretical considerations by Lambert, and comparisons with standard trailing arm air suspensions
revealed that the system had significant merit in overcoming air suspension deficiencies that had been
described by drivers of “problem” prime movers as described in Lambert 2000.
“Problem vehicles” exhibit all or nearly all of:
• excessive whole body vibration at the drivers seat,
• excessive maintenance due to vibration related damage in the broad sense, c
• onstant steering input to control wandering, and
• the occurrence of significant divergences from the intended vehicle path without driver input (darting).
A risk assessment was undertaken and tests carried out that determined there were no increased operational
risks with the new suspension, and that showed how other risks would be minimised.
Whole body vibration, excessive maintenance, wandering and darting have disappeared with fitting of the
modified suspension; traction is comparable to mechanical six-rod suspensions; tyre wear is significantly
reduced; and brake pad wear is even.
This led to further consideration of the theoretical reasons for these improvements, and these are discussed.
Comments are made about areas of poor design with other air suspensions, and recommendations made in
respect of air suspension design. One recommendation arising from this evaluation is that air suspensions,
whose design requires two height control vales to maintain the vehicle upright, should be banned on safety
grounds. They result in a much increased likelihood of “problem vehicle” behaviour as well as loss of
control and roll over, particularly on highly cambered roads.
INTRODUCTION
In Australia there has been a massive move to air suspensions on heavy vehicles, largely brought about by
the introduction in most states of extra axle mass for vehicles with road friendly air suspensions.
Prior to that time most suspensions were mechanical spring designs with load sharing functionality built in.
Two mechanical spring designs are shown schematically below.
The load sharing mechanisms provided for high speed load transfer and hence at speed provided dynamic
load sharing. With these suspensions most of the damping was achieved through interleaf friction combined
with tyre hysteresis.
Proceedings 8th International Symposium on Heavy Vehicle Weights and Dimensions
‘Loads, Roads and the Information Highway’
Proceedings Produced by: Document Transformation Technologies
14th - 18th March, Johannesburg, South Africa
ISBN Number: 1-920-01730-5
Conference Organised by: Conference Planners
Figure 1. Trailer style bogie suspension with load sharing through a rocker arm.
Figure 2. Bogie suspension with load sharing through a pivoted arm.
Air suspensions had been trialed on heavy freight vehicles in the 1970’s but were abandoned for a number of
reasons including the fact that freight vehicles tended to exhibit excessive roll from side to side.
During the 1980’s and into the 1990’s testing was undertaken to “prove” that new air suspension designs
fitted with appropriate shock absorbers reduced pavement damage. For example Sweatman 1982. Much of
the testing was done using pavement testing machinery similar to that shown below.
Almost all this equipment uses at most dual wheels on a suspension unit, and the majority of testing is at low
speeds of 50 km/h or less. Units are of designs that generally do not allow the testing of complete axle group
suspensions.
As far as the author can ascertain, in all the road friendly suspension testing there was only a limited amount
of testing at 80 km/h and virtually no testing above this speed.
Hence there arises a question about whether air suspensions are road friendly at the speeds most trucks travel
at on open roads. That is speeds of 90 km/h or greater.
Figure 3. Pavement testing machines.
PROBLEM HEAVY VEHICLES
With the introduction of air suspended heavy vehicles, there was an apparent increase in the frequency of
“problem” heavy vehicles.
“Problem heavy vehicles” exhibit all or nearly all of:
• excessive whole body vibration at the drivers seat,
• excessive maintenance due to vibration related damage in the broad sense,
• constant steering input to control wandering, and
• the occurrence of significant divergence from the intended vehicle path without driver input (darting).
The problems may be described as those that arise from chaotic behaviour of the prime mover – that is
non-uniform and/or random periodic movements of the major components of the prime mover in relation to
each other and the road pavement at frequencies ranging from less than 1 Hz to 70 Hz or greater.
These vehicles exhibit high levels of what can be broadly described as “vibration”
“Problem” heavy vehicles were primarily those with light torsionally flexible long chassis; raised heavy
cabins and sleepers; large, high, and heavy high powered motors; torsionally rigid trailers carrying high COG
loads at axle mass limits; and were vehicles which travelled on secondary highways at 100 km/h.
Owners of these vehicles commonly reported that the vehicles were fine up to 80 km/h. Lambert 2000
discusses these problem vehicles.
Work by the authors and Bill Haire of Haire Truck and Bus Repairs has shown that these problems can be
eliminated by changes to the drive axle air suspension and corrections to poorly designed transmission shaft
layouts.
This again raised questions about the current design of air suspensions as fitted to heavy vehicles.
FORS HEAVY VEHICLE SAFETY PROJECT
This project attempted to determine the cause of the behaviour of the problem vehicles, including
“wandering, poor handling, instability and abnormal vibration.” Eight trucks were heavily instrumented to
record a large number of variables including air bag pressures.
Unfortunately only the rear drive axle air bag pressures were monitored.
Vehicles were operated on three sections of road as close to 100 km/h as was possible. The same test driver
was used and he chose a speed that he was comfortable with given the vehicle’s handling and the road
environment. And all vehicles towed the same flat bed trailer and the same load.
There were five “problem” vehicles tested (File F1, F3, F4, F6 and F26) and three bench mark vehicle
supplied by manufacturers (BM1, BM2 and BM3).
As shown in Table 1 below and highlighted in yellow, mean air bag pressures as measured in the left and
right airbags on the rear drive axle varied dramatically on three of the eight vehicles. In the worse case the
ratio of airbag pressures was around 2:1.
Table 1. FORS heavy vehicle safety report – air bag pressure data.
File
Vehicle
F1
F3
F4
F6
F26
BM1
BM2
BM3
Kenworth T900
Kenworth C501
Mack CH Fleetliner
Ford LT 9513
Mack CH Elite
Mack CH Fleetliner
Kenworth T900
Volvo NH
Euroa (straight section)
LH air bag
RH air bag
pressure (psi)
pressure (psi)
mean
mean
59.5
59.7
85.2
37.1
49.6
53.5
74.6
54.4
53.3
54.0
53.7
53.1
69.6
51.4
56.5
54.0
Nagambie (winding) section
LH air bag
RH air bag
pressure (psi)
pressure (psi)
mean
mean
59.3
59.6
69.3
40.8
49.3
53.1
72.5
54.9
52.9
53.6
53.1
52.4
68.3
52.5
51.0
57.5
It should be noted that mean speeds were not always close to 100 km/h. As shown in Table 2 below, mean
speeds varied significantly. On the straight section one vehicle had an average speed of only 84.7 km/h. And
on the hilly section while two vehicles were able to maintain mean speeds close to 100 km/h, the rest had
means that were much lower.
On Australian secondary highways there is generally a significant degree of camber present. As vehicles
travel on the left hand side of the road, this camber will cause the mass on the left hand tyres to be around
5%-10% and up to 18% greater than that on the right hand tyres.
Analysis of moments and forces requires that for the same location of the vehicle’s centre of mass, the load
through the tyres remains the same, even if the air bag pressures vary dramatically. Hence the forces
generated by the unequal air bag pressures above are largely redistributed through torsion in the axle, and the
net effect of the unequal pressures is to distort the suspension, Due to roll steer effects on trailing arm air bag
suspensions, the end result is that the axle is effectively dynamically misaligned leading to all the problems
associated with a misaligned drive suspension.
Table 2. FORS heavy vehicle safety report – speed data.
File
Vehicle
F1
F3
F4
F6
F26
BM1
BM2
BM3
Kenworth T900
Kenworth C501
Mack CH Fleetliner
Ford LT 9513
Mack CH Elite
Mack CH Fleetliner
Kenworth T900
Volvo NH
Speed Euroa section
mean (km/h)
96.4
Na
Na
98.5
97.9
84.7
92.8
96.5
Speed Nagambie section
mean (km/h)
96.4
88.1
Na
95.6
83.8
88.7
91.6
88.9
Modelling by Lambert of airbag pressures versus forces and the mass distribution across the vehicle axle
groups also suggests that the average airbag pressures at the rear drive axle may be up to 10% higher than
expected for the load carried. This will have the effect of transferring load onto the steer axle, a phenomenon
reported by the drivers of these vehicles as “it feels like the rear air bags have jacked up.”
But what causes this unequal pressure?
CURRENT AIR SUSPENSION DESIGNS
A typical design is shown schematically below.
Air bag
internal
height =
150 - 250
mm
Axle
movement
H
Air bag
height
change =
1.5 H - 2.0
H
Figure 4. Typical air suspension – shock absorbers not shown.
Because of the geometry of these units, the air bag height changes by 1.5 – 2.0 times the change in height of
the axle.
Typically the connections between the air bags are small-bore tubes with sharp edges orifices into the
airbags.
However airbag suspensions also incorporate a height control valve (HCV) or valves aimed at ensuring the
correct ride height of the vehicle so that geometry of the vehicle is maintained regardless of the load. The
basic arrangement is shown below.
The HCV is actuated typically by a rod from the valve lever to the rear of the trailing arm on the rear axle.
When a load is placed on a truck the trailing arm rises and the HCV acts to pressurise the air bags until the
vehicle height returns to its set height. Conversely when a load is removed the HCV acts to exhaust air from
the air bag to atmosphere.
System air pressure
Height control valve
Figure 5. Bogie air suspension with height control valve – neutral position.
However the HCV also acts during travel as the vehicle travels over roads that are uneven. As shown below
when the rear axle goes over a bump the HCV acts to put air into the rear air bag. Because the tube length to
the rear airbag is small, air will flow more freely into this bag compared to the front air bag that has a much
longer tube length.
System air pressure
Height control valve
When axle raises the HCV acts
to put more air into the airbag to
maintain/restore height.
Figure 6. Bogie air suspension with height control valve – rear axle raised.
And if the rear axle drops into a depression the HCV acts to exhaust air from the rear airbag. And the air will
move more freely out of the rear air bag.
The net effect of the movement of the HCV with the vehicle travelling at speed depends on the sensitivity
and design of the HCV, the location of the HCV, the system air pressure versus the airbag pressure, and the
geometry of the tubes connecting the airbags. Unless all connections to the airbags are of equal length and
with the same air line fittings, there will be a tendency for the air bag pressures to change until the net flow
into and out of all the airbags is the same.
With some suspension designs with a lot of compliance, two HCV’s are used to maintain the truck upright.
Where two HCV’s are used, one on each side, the effect of road camber is to cause the vehicle to lean in the
direction of the camber (toward the low side). In Australia, with sensitive HCV’s the left bag will be
pressurised and the right bag depressurised. This is the case with all the vehicles above where there is a large
difference in air bag pressures from left to right.
In summary, the HCV changes the suspension from a passive suspension to an active suspension, and that
can lead to significant problems in handling as a result of significant variations in air bag pressures.
CURRENT AIR SUSPENSION DESIGNS – DYNAMIC LOAD SHARING AND DAMPING
In order for airbag suspensions to load share, air must be transferred from airbag to airbag with each bump or
depression. However the small bore size of almost all current designs prevents the transfer of large amounts
of air.
Airbag
Sharp edged nozzle
restricts flow
Figure 7. Effect of sharp edged tube connection to the air bag.
Further connections of these tubes to the airbags generally generate sharp edged orifices as shown above.
These act to further restrict flow, possibly by as much as 50%.
As a result in a high speed environment the airbag acts as though it is blanked off and there is no dynamic
load sharing.
The flow of air to and from an airbag generates turbulence, and hence energy loss as heat. This energy loss
acts to dampen oscillations. The concept is shown below.
AIR DAMPING CONCEPT - DAMPING VERSUS
PIPE TO BAG AREA RATIO
DAMPING
0.6
0.5
0.4
0.3
Low pressure loss
= low damping
0.2
0.1
0
Low flow =
0
0.1
low damping
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
AREA RATIO - PIPE AREA/ BAG AREA
Figure 8. Concept of air suspension internal damping.
For small bore tubing there is little airflow, hence the internal damping effect is low. Conversely for very
large tubes there is little turbulence generated because there is only a low pressure drop. Hence at some
intermediate point maximum damping through air bag to air bag flow is achieved.
INNOVATIVE DYNAMIC LOAD SHARING AIR SUSPENSION
Haire Truck and Bus Repairs over the past 10 or more years has developed an air suspension system that
tends to correct all these adverse behaviours.
The system consists of:
• Large diameter (50.8 mm) orifice transmission lines with biased airflow control to achieve fast response
and optimal load sharing between axles in an axle group,
• Relatively large diameter pipes (20 mm) from the air bag to the transmission lines;
• Detailed design features that generate around 7% damping through internal air flow; and
• Innovative mean ride height feedback linkage to generate correct feedback signal without the HCV
actuating with every bump or depression
Figure 9. Large diameter orifice transmission line.
Figure 9 shows the fittings for one side of a tri-axle group. An appropriate length of 50.8 mm hose is fitted
between the fittings shown.
Small bore tubing ~ 5 mm
ID from height control valve
to large diameter hose
Low flow height
Control valve
Steel rod connected to and
spanning gap between axles or
differentials
Figure 10. Mean ride height feedback linkage.
A system of rods is used to sense the mean height of a bogie axle group as shown in Figure 10. For a tri-axle
group sensing the centre of the centre axle may suffice. As a result of the limited degree of actuation of the
HCV air consumption is dramatically reduced.
In McLean et al (2003) the change in airbag pressure in response to a step input was calculated with the
results shown below. The figures assume both wheels on an axle negotiate the same 100 mm step at the same
time.
Figure 11. Details of the Haire fast response air suspension system.
Figure 12. Airbag pressures in response to a step change in height for 3 suspensions.
The average increase in “road damage” arising from the first pressure spikes can be calculated from the
increase on pressure and determined for a 4th power rule and a 12th power rule.
Table 3. Pressures and “road damage” for three suspensions.
Suspension
Standard capillary bogie
Fast response system bogie
Fast response system triaxle
Average
pressure
(kPa)
485
485
323
Maximum
pressure
(kPa)
641
540
347
“Road damage” 1st
half cycle – 4th power
2.19
1.32
1.21
“Road damage” 1st
half cycle – 12th
power
13.28
2.43
1.80
The much greater potential road damage of the standard capillary tube suspension is readily apparent from
the figures above.
And the more uniform pressure on the road is also apparent with the increase in pressure for the capillary
tube suspension being 32.2% whereas for the two other suspensions the increases are 11.2% and 7.5%.
RISK ASSESSMENT
Of concern with this new suspension was the vehicle response with operational and worn out shock
absorbers, and the response to an airbag failure.
Tests of simulated airbag failures were undertaken at speed around a roundabout and at 100 km/h in straight
ahead travel. These showed that the vehicle dropped uniformly onto the bump stops and that there was no
unusual or difficult to control vehicle behaviour. This is because the loss of air on the “burst” airbag side is
so rapid that, as the airbags on the other side cannot then support the load they collapse almost as rapidly.
Similarly the removal of shock absorbers created no problems. Around 10% of damping comes from the
suspension bushes and the tyres, and 7% from internal damping. So the balance of the required 20% damping
for road friendly air suspensions from the shock absorbers is only 3% so there is little effect on performance.
And this low requirement for damping from shock absorbers means that light single acting shock absorbers
can be used for these suspensions.
CONCLUSION
In spite of testing in the 1980’s and later showing that air suspensions are “road friendly” there is now
serious doubts as to whether this is true in high speed situations (> 80 km/h). This is because current designs
are not dynamically load sharing, and the design of height control systems is such that the suspension airbag
pressures can vary significantly and help to generate chaotic behaviour in heavy trucks.
Theoretical considerations show that this is primarily due to the design of the HCV system including the
small bore tube plumbing connecting the airbags.
An innovative modification to appropriate air suspensions results in these suspensions becoming
dynamically load sharing. Theoretical modelling shows that the pressure rise in the airbags in the fast
response system is about a third of that in the traditional system.
Though not done in this paper the same theoretical analysis can be used to demonstrate why twin HCV
systems generate large differences in airbag pressures from side to side. These pressure variations are a
significant cause of “problem vehicle” behaviour in heavy trucks and represents a safety hazard. Hence
consideration should be given to reviewing the design of this style of suspension to avoid these deficiencies.
REFERENCES
1. Lambert J M, November 2000. Big prime movers – why does australia have serious problems? And how
do we fix them? Paper presented at the Road Safety Research, Policing and Education Conference, 26-28
November 2000, Brisbane Australia.
2. McLean A G, and Li B; Quasi-static Modelling and Simulation of an Innovative Fast Response Heavy
Vehicle Air Suspension System, Paper presented at the 2003 ARRB Transport Research Conference.
3. Sweatman P and McFarlane S, Investigation into the Specifications of Heavy Trucks and consequent
Effects on Truck Dynamics and Drivers: Final Report; FORS (now ATSB) 2000.
4. Sweatman, P. F. (1983). “A study of dynamic wheel forces in axle group suspensions of heavy vehicles,”
Rep. No. 27. Australian Road Research Board, Melbourne.